WO3 slab模型scf收敛判断

rabbitkiller

大家好

最近尝试用cp2k算一个WO3的slab模型，&DIAGONALIZATION里面，ALGORITHM用STANDARD无法让scf收敛，尝试用了DAVIDSON方法，出现了神奇的一幕：

SCF WAVEFUNCTION OPTIMIZATION

  Step     Update method      Time    Convergence         Total energy    Change
  ------------------------------------------------------------------------------
  POISSON| Poisson solver converged in 1 iteration.
     1 NoMix/Dav.  0.40E+00   14.1     1.93787018    -10893.8156308417 -1.09E+04
  POISSON| Poisson solver converged in 1 iteration.
     2 Broy./Dav.  0.40E+00   10.2    11.49045924    -10758.8227421560  1.35E+02
  POISSON| Poisson solver converged in 1 iteration.
     3 Broy./Dav.  0.40E+00   10.2    10.46847170    -10578.0479201526  1.81E+02
  POISSON| Poisson solver converged in 1 iteration.
     4 Broy./Dav.  0.40E+00   10.2     5.59099113    -10730.9251099218 -1.53E+02
  POISSON| Poisson solver converged in 1 iteration.
     5 Broy./Dav.  0.40E+00   10.3     0.42627641    -10899.9810917096 -1.69E+02
  POISSON| Poisson solver converged in 1 iteration.
     6 Broy./Dav.  0.40E+00   10.3     1.69147304    -11052.8641341044 -1.53E+02
  POISSON| Poisson solver converged in 1 iteration.
     7 Broy./Dav.  0.40E+00   10.3     1.32338206    -11163.3699094426 -1.11E+02
  POISSON| Poisson solver converged in 1 iteration.
     8 Broy./Dav.  0.40E+00   10.4     0.87180172    -11201.9004837754 -3.85E+01
  POISSON| Poisson solver converged in 1 iteration.
     9 Broy./Dav.  0.40E+00   10.4     1.44385395    -11197.9567093136  3.94E+00
  POISSON| Poisson solver converged in 1 iteration.
    10 Broy./Dav.  0.40E+00   10.5     1.80357315    -11190.5487526823  7.41E+00
  POISSON| Poisson solver converged in 1 iteration.
    11 Broy./Dav.  0.40E+00   10.5     1.78527179    -11189.7848723211  7.64E-01
  POISSON| Poisson solver converged in 1 iteration.
    12 Broy./Dav.  0.40E+00   10.5     1.29607565    -11222.0507220931 -3.23E+01
  POISSON| Poisson solver converged in 1 iteration.
    13 Broy./Dav.  0.40E+00   10.5     1.90950247    -11205.3665241203  1.67E+01
  POISSON| Poisson solver converged in 1 iteration.
    14 Broy./Dav.  0.40E+00   10.5     1.78003821    -11180.9128705079  2.45E+01
  POISSON| Poisson solver converged in 1 iteration.
    15 Broy./Dav.  0.40E+00   10.5     0.47105641    -11168.1744067137  1.27E+01
  POISSON| Poisson solver converged in 1 iteration.
    16 Broy./Dav.  0.40E+00   10.5     0.70553900    -11164.5328141507  3.64E+00
  POISSON| Poisson solver converged in 1 iteration.
    17 Broy./Dav.  0.40E+00   10.5     1.81922711    -11167.2807185098 -2.75E+00
  POISSON| Poisson solver converged in 1 iteration.
    18 Broy./Dav.  0.40E+00   10.5     1.78067717    -11177.8700988120 -1.06E+01
  POISSON| Poisson solver converged in 1 iteration.
    19 Broy./Dav.  0.40E+00   10.5     1.19231111    -11194.6481976790 -1.68E+01
  POISSON| Poisson solver converged in 1 iteration.
    20 Broy./Dav.  0.40E+00   16.3     1.75890726    -11195.3644120988 -7.16E-01
  POISSON| Poisson solver converged in 1 iteration.
    21 Broy./Dav.  0.40E+00   10.5     1.82585112    -11169.2912594905  2.61E+01
  POISSON| Poisson solver converged in 1 iteration.
    22 Broy./Dav.  0.40E+00   10.5     1.81789781    -11188.2772906849 -1.90E+01
  POISSON| Poisson solver converged in 1 iteration.
    23 Broy./Dav.  0.40E+00   11.4     1.89034157    -11187.8785538563  3.99E-01
  POISSON| Poisson solver converged in 1 iteration.
    24 Broy./Dav.  0.40E+00   11.7     1.48829146    -11432.8928813799 -2.45E+02
  POISSON| Poisson solver converged in 1 iteration.
    25 Broy./Dav.  0.40E+00   11.3     1.70554652    -11406.3661456862  2.65E+01
  POISSON| Poisson solver converged in 1 iteration.
    26 Broy./Dav.  0.40E+00   10.5     0.20119363    -12615.6648838136 -1.21E+03
  POISSON| Poisson solver converged in 1 iteration.
    27 Broy./Dav.  0.40E+00   10.8     1.63762659    -12617.0960578040 -1.43E+00
  POISSON| Poisson solver converged in 1 iteration.
    28 Broy./Dav.  0.40E+00   10.8     1.69812030    -12569.0102432443  4.81E+01
  POISSON| Poisson solver converged in 1 iteration.
    29 Broy./Dav.  0.40E+00   10.8     1.75467632    -12579.9263952916 -1.09E+01
  POISSON| Poisson solver converged in 1 iteration.
    30 Broy./Dav.  0.40E+00   10.6     1.95426159    -12553.9918582203  2.59E+01
  POISSON| Poisson solver converged in 1 iteration.
    31 Broy./Dav.  0.40E+00   10.6     1.75567134    -14110.1762008493 -1.56E+03
  POISSON| Poisson solver converged in 1 iteration.
    32 Broy./Dav.  0.40E+00    9.0     1.69536393    -12882.0530608784  1.23E+03
  POISSON| Poisson solver converged in 1 iteration.
    33 Broy./Dav.  0.40E+00    7.1     0.0000E+00    -14218.4234146328 -1.34E+03

  *** SCF run converged in    33 steps ***


请问这个能算是scf收敛么？我觉得不像，毕竟总能量还在很厉害地抖动，但convergence居然变成0了。

想问问大家看这是怎么回事？

谢谢！

sobereva

像是触发了bug，可以向官方google group报告